Evaluating Hospital Nurse Resources as a Driver of Outcomes Differences in Hospitals Serving High-Risk Patients

University Of Pennsylvania, Philadelphia PA

Investigators

Linked publications & trials

Abstract

: The intractable issue of differences in care quality across hospitals has been described for decades and most recently came to the forefront of public attention with differences in hospital COVID-19 mortality. Despite multiple studies which attribute a large share of outcome differences to quality differences in hospitals where significant proportions of high-risk patients are treated (i.e., high-risk-serving hospitals), little is known about what modifiable factors underlie poorer quality care. This proposal takes a multilevel perspective to identify the contributions of individual socio-demographic and hospital factors to identify modifiable factors that can be targeted through ‘upstream’ interventions to achieve reliably high-quality hospital care for all patients. We hypothesize that differences in hospital outcomes are due, in large part, to differences in the modifiable nurse resources of hospitals—with fewer nurse resources in some hospitals. This uneven distribution of nurse resources is a vestige of historic underinvestment that continues to pervade hospital care. We focus on nurse resources, since having enough nursing staff to deliver timely and effective care, a favorable work environment in which nurses have clinical autonomy in their practice and strong interdisciplinary teamwork, a skill mix rich in registered nurses, and high proportions of bachelors-prepared nurses and advanced practice nurses, have all been associated with better patient outcomes, particularly for high-risk patients. In this study, we evaluate the impact of hospital-level differences in nurse resources on patient outcomes, including in-hospital and 30-day mortality, readmission, and hospital length of stay. This observational study of over 900,000 older adult patients in nearly 250 hospitals investigates (1) whether differences in nurse resources between high-risk-serving and other hospitals explain outcome differences; (2) whether the outcomes-advantages of having superior hospital nurse resources are enhanced in the presence of other hospital characteristics, including for example, physician staffing, greater numbers of APRNs, or teaching hospitals; and (3) estimates the improvements in patient outcomes, such as lives saved, that could be expected if nurse resources in high-risk-serving hospitals were similar to other hospitals. Nurse resources are measured using survey data from over 16,000 nurses in nearly 250 hospitals to describe multiple aspects of the clinical nurse resources. Using a unique hospital identifier, nurse responses will be linked with Medicare patient records, and socio-demographic factors. Our analytic approach uses multi-level nested (hierarchically-related) linear and logistic regression models (with interaction terms) to accomplish our aims. If our hypotheses are confirmed, the findings will add evidence to inform high-impact actionable ‘upstream’ solutions to achieve reliably high-quality hospital care and patient outcomes, by leveraging the most abundant patient care resource already existing in every hospital—nurses.